Synergistic effect of trivalent (Gd³⁺, Sm³⁺) and high-valent (Ti⁴⁺) co-doping on antiferromagnetic YFeO₃

Abstract

Monophasic polycrystalline powders of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) were successfully synthesized via a low temperature solid-state synthesis route. The X-ray diffraction and Raman spectroscopy studies indicate that all the calcined powders with R³⁺ (Gd³⁺, Sm³⁺) at Y³⁺ and Ti⁴⁺ at Fe³⁺ sites were crystallized in an orthorhombic phase associated with a change in lattice parameters. The Williamson–Hall method employed to calculate the strain revealed that the strain increased with the increased concentration of dopants ((Gd³⁺, Sm³⁺) at Y³⁺) compared to an increase in the size of crystallites, corroborating the findings of SEM. Analysis of diffuse reflectance spectra indicated a drop in bandgap from 1.93 eV to 1.86 eV and 1.96 eV to 1.91 eV for Gd, Ti co-doping and Sm, Ti co-doping respectively, demonstrating the capacity of the synthesized powders to absorb visible light. Absorbance spectra also revealed the existence of mixed states of Fe³⁺ and Fe⁴⁺ which was corroborated by XPS studies. The magnetic hysteresis loop analysis at room temperature illustrated that with co-doping, there is a strong enhancement in magnetization as well as coercivity, suggesting a strong transition from anti-ferromagnetic behaviour to ferromagnetic behaviour. Pertaining to the greatly improved optical and magnetic properties with the addition of (Gd³⁺, Sm³⁺) at Y³⁺ and Ti⁴⁺ at Fe³⁺ sites, these materials are anticipated to be of potential use in various applications.

---

Introduction

The existence of various physical properties associated with a single phase material has resulted in the study of and technological interest in multiferroic materials.¹ These multiferroic materials could be employed for several applications that include temperature induced spin switching,² magnetic field induced spin switching,^3,4 solid oxide fuel cells, photo-catalysis, magneto-optical devices, capacitors, magneto-electric sensors in transducers, microwave electronics, and spintronics, owing to their magnetic and electric polarization effects.^5–7 RFeO₃ has a distorted perovskite structure that crystallizes in the orthorhombic Pmna space group.⁸ The crystallographic unit cell contains 4 equivalent iron (Fe) ions. Although RFeO₃ has a distorted structure, FeO₆ remains essentially an octahedron with six oxygen anions surrounding the Fe-ion, leading to crystal field splitting. The typical rare-earth orthoferrite YFeO₃ falls into this category. There have been several reports on the dielectric and magnetic properties of YFeO₃, even though it did not exhibit conventional ferroelectric effects at room temperature due to its low Curie temperature (−256 °C). Nevertheless, the narrow energy bandgap (1.8–2.6 eV) makes YFeO₃ a viable material for magneto-optical applications.⁹ The weak ferromagnetic behaviour of YFeO₃ is due to the canting of Fe³⁺ ions in an orthorhombic structure. However, it was demonstrated that the magnetic properties of YFeO₃ can be enhanced by means of doping. For instance, the non-magnetic Y³⁺ is doped by divalent¹⁰ and trivalent ions^11,12 and Fe³⁺ by trivalent¹³ and high valency ions.¹⁴ Doping at Y³⁺ resulted in the improvement of magnetization while the Fe³⁺ doping enhanced the coercivity. To the best of our knowledge, there exist no reports in the literature on co-doping of rare-earth (Gd³⁺, Sm³⁺) at Y³⁺ and Ti⁴⁺ at Fe³⁺. Therefore the objective of the current report is to investigate into the effect of co-doping of rare-earth (Gd³⁺, Sm³⁺) and Ti⁴⁺ on the structural, magnetic and optical properties of YFeO₃.

There are several synthesis routes that are known^15–19 to obtain the polycrystalline powders of Y_1−xR_xFe_1−(4/3)yTi_yO₃ where R = Gd, Sm. These multiple synthesis routes include the most popular solid-state and sol–gel routes. However, there are certain disadvantages associated with the solid-state synthesis route. First being the formation of secondary phases such as Y₂O₃ and Y₃Fe₅O₁₂. Secondly, the existence of a hexagonal phase at low temperatures, which at higher temperatures becomes an orthorhombic phase. Our previous research identified the effective preparation of polycrystalline YFeO₃ by sol–gel synthesis route in detail.¹² Sol–gel approach is beneficial relative to the solid-state system because it allows for homogeneous mixing of the precursor ions at a molecular level. However, other drawbacks associated with this process are (i) the creation of pores due to the generation and trapping of gases, (ii) the coexistence of hexagonal and orthorhombic phases which could be solved by precisely regulating the calcination temperatures and (iii) usage of certain toxic complexing agents.

In view of the aforementioned disadvantages, an easy and efficient method that is slightly different from the above method has been developed. It is a low temperature solid-state route for the synthesis of Y_1−xR_xFe_1−(4/3)yTi_yO₃ which has the advantage over other methods in phase purity, precision in phase control, low-cost and ultrafine particle size. This synthetic method was effectively employed to obtain pristine YFeO₃.¹³

Materials and methods

The polycrystalline powders of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) have been successfully obtained using low temperature solid-state synthesis method. The precursors used were Y(NO₃)₃·6H₂O (AR grade, purity > 98%), Fe(NO₃)₃·9H₂O (AR grade, purity > 98%), Sm(NO₃)₃·6H₂O (Sigma-Aldrich, purity > 99%), Gd(NO₃)₃·6H₂O (Sigma-Aldrich, purity > 99%), TiO₂ (Sigma-Aldrich, purity > 99%) and citric acid (AR grade, purity > 99%). Initially, the relevant precursors were weighed according to the stoichiometric ratios and were ground in an agate mortar for 30 min. This initial grinding resulted in the formation of light brown mush implicating the complex formation. This mush was then heated at 120 °C for 2 h which facilitated the removal of free water and yielded dark brown powder. These powders were then heated at 350 °C for 6 h to aid the complete decomposition of nitrates. Subsequently, these were finely ground and subjected to heating at 900 °C for 1 h resulting in the formation of monophasic polycrystalline powders. The schematic is given in Fig. 1.

Fig. 1 Schematic of low temperature solid-state synthesis route.

Mechanism of low temperature solid-state synthesis

Solid-state reaction involves four stages (i) diffusion, (ii) reaction, (iii) nucleation and (iv) growth. Diffusion takes place when the reactant or precursor molecules are in contact, followed by chemical reaction yielding the product. The nucleation is completely achieved until the product molecules aggregate to a certain scale. Once the product forms an independent crystal phase, crystal nuclei expand to a certain size. In the low temperature solid-state reaction, the reactants are brought into full contact as they reach a micron-level mixture. This is achieved by grinding the reactants which reduces their size aiding the surface energy to increase. As some of the reactants are hygroscopic, several micron liquid baths are created during the grinding process. This liquid environment helps the surface molecules of the reactants to aggregate and then react rapidly with one another. This results in the reaction to occur faster. The citric acid apart from acting as a chelating or complexing agent, provides the energy in the form of heat during calcination helping Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) to form at low temperatures.

Characterization

The polycrystalline powders were characterized using X-ray powder diffractometer (X'Pert Pro, Panalytical) with Cu K_α = 1.5408 Å. The average crystallite size and lattice strain of these samples have been determined using Williamson–Hall plots. The structural analysis was further strengthened by recording Raman spectra using Raman microscope, (Thermo Scientific, laser excitation, λ = 780 nm). Microstructural and elemental analyses were performed using Jeol IT300 scanning electron microscope (SEM). Magnetic hysteresis loops at room temperature were obtained using vibrating sample magnetometer (Model EZ9, MicroSense, USA). The bandgap of the polycrystalline powders of the different formulations were measured using diffuse reflectance spectroscopy (DRS) in the range 200–800 nm employing UV-Vis spectrophotometer (Perkin Elmer LS 55). The X-ray photoelectron spectroscopic studies for all the samples were pursued by Fison Instruments S-probe TM2803.

Results and discussion

The room temperature X-ray powder diffraction (XRD) patterns of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) are represented in Fig. 2. XRD studies confirmed the formation of monophasic crystalline orthorhombic phase for all the samples under investigation with no detectable impurity peaks. This outcome from XRD studies indicate that the reaction is complete and the products formed are of required purity. The partial substitution of Gd³⁺ (x = 0.05, 0.10, 0.15), Sm³⁺ (x = 0.05, 0.10, 0.15) at Y³⁺ sites and Ti⁴⁺ (y = 0.05) at Fe³⁺ sites could be evidenced from the downward shift in the peak positions with the increase in dopant concentration, as shown in the insets of Fig. 2. The downward shift in the peak positions with an increase in dopant concentration implies that there is an increase in lattice parameters. For orthorhombic system, the peak positions and lattice parameters are related by (h/a)² + (k/b)² + (l/c)² = 4 sin² θ/λ², where h, k, l are Miller indices; a, b, c are lattice parameters; θ is the peak position; λ is the wavelength of the X-rays used. Further, as we are doping Gd³⁺ (x = 0.05, 0.10, 0.15), Sm³⁺ (x = 0.05, 0.10, 0.15) at Y³⁺ sites whose ionic radii (Gd³⁺ = 1.24 Å, Sm³⁺ = 1.10 Å) are relatively higher than that of Y³⁺ (1.06 Å), such shifts are expected. Moreover, if the partial substitution of Gd³⁺, Sm³⁺ and Ti⁴⁺ deviates from Y³⁺ and Fe³⁺ respectively, then the secondary phases/impurities must be formed which will reflect as extra peaks in XRD pattern. The absence of detectable X-ray peaks which are unaccounted for in these patterns confirm the effective doping at appropriate sites.

Fig. 2 XRD patterns obtained for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

The computed lattice parameters based on XRD data are depicted in Table 1. The values increased with the increase in the rare earth (Gd, Sm) doping concentration at Y³⁺ and decreased with Ti⁴⁺ doping at Fe³⁺ suggesting the effectiveness of doping of Gd³⁺, Sm³⁺ which results in the increase in YFeO₃ (r_Sm³⁺ (1.24 Å) > r_Gd³⁺ (1.10 Å) > r_Y³⁺ (1.06 Å)) lattice volume, whereas the substitution of Ti⁴⁺ at Fe³⁺ results in the reduction of YFeO₃ (r_Fe³⁺ (0.645 Å) > r_Ti⁴⁺ (0.605 Å)) lattice volume.²⁰ This incessant evolution of lattice parameters also indicate the successful doping of Gd³⁺, Sm³⁺ of different concentrations at Y³⁺ sites and Ti⁴⁺ (0.05) at Fe³⁺ sites in YFeO₃.

Table 1 Refinement and lattice parameters of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05). Characterization tool: XRD

Name of the compound	Dopant concentration	Lattice parameters			Agreement factors
		a (Å)	b (Å)	c (Å)	Rwp (%)	Rp (%)	GOF (%)
YSFTO	x = 0.05, y = 0.05	5.589	7.603	5.283	2.73	1.96	1.37
	x = 0.10, y = 0.05	5.593	7.619	5.291	3.07	2.24	1.58
	x = 0.15, y = 0.05	5.597	7.626	5.307	3.27	2.35	1.61
YGFTO	x = 0.05, y = 0.05	5.591	7.613	5.287	3.03	2.15	1.57
	x = 0.10, y = 0.05	5.594	7.618	5.294	2.88	2.08	1.64
	x = 0.15, y = 0.05	5.601	7.622	5.301	3.26	2.13	1.94

---

The X-ray diffraction peaks for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) powders were indexed to YFeO₃ orthorhombic crystal structure (ISCD 98-008-0866) associated with ‘Pmna’ space group of ‘mmm’ point group. The perovskite (ABO₃) structure incorporating different elements in the periodic table as structural distortion, though ideally it is cubic. These distortions are essentially of three types, tilting of BO₆ octahedra, polar cationic displacement and Jahn–Teller distortions. Out of these three, tilting of BO₆ octahedra is most commonly found in Pmna space group to which YFeO₃ belongs. The increase in lattice parameters with an increase in dopant concentration leads to an increase in structural distortion. The structural distortion could be calculated using O'Keefe geometric approximation.²¹

where θ₁, θ₂ and φ₁, φ₂ are bond angles and octahedral tilt angles respectively.

The sharing of O²⁻ ions between two FeO₆ octahedral structures leads to the formation of two kinds of Fe–O–Fe superexchange bonds, θ₁ and θ₂. These bond angles and tilt angles obtained using crystallographic information data acquired from Reitveld analysis were represented in VESTA software (ver. 3.5.2). The extent of distortion with respect to doping concentration i.e., the change in tilt angles and bond lengths are given in Table 2.

Table 2 Bond angles Fe–O–Fe (θ₁, θ₂), bond lengths (Fe–O) and octahedral tilt angles (φ₁, φ₂) of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05). Characterization tool: XRD

Name of the compound	Bond angles		Tilt angles		Bond length (Å)
	θ1	θ2	φ1	φ2	(Fe/Ti–O)
Y1−xSmxFe1−(4/3)yTiyO3
x = 0.05, y = 0.05	144.615	143.59	21.35	22.49	2.032
x = 0.10, y = 0.05	144.613	143.582	21.343	22.50	2.035
x = 0.15, y = 0.05	144.607	143.570	20.863	22.51	2.037
Y1−xGdxFe1−(4/3)yTiyO3
x = 0.05, y = 0.05	144.831	143.712	21.22	22.42	2.03
x = 0.10, y = 0.05	144.829	143.689	21.21	22.44	2.034
x = 0.15, y = 0.05	144.597	143.560	21.42	22.51	2.044

---

With the doping at Y³⁺ and Fe³⁺ sites the structure should remain as orthorhombic. The phase stability is measured using Goldschmidt tolerance factor which is given by

where R^eff_A = R_Y(1 − x) + R_Sm/Gd(x), R^eff_B = R_Fe[1 − (4/3)y] + R_Ti(y)

Crystallite size and strain analysis

The crystallite size could be calculated using Scherrer formula associated with the full width at half maximum of the peaks as given in eqn (1).

The Bragg peak width of the obtained XRD patterns also consists of instrument broadening effects. To circumvent such aberrations, the diffraction pattern obtained from the standard silicon sample was used to ascertain the instrument broadening. The relation given in eqn (2) is used to obtain the instrument broadening corresponding to the diffraction peaks of the samples under investigation.

β_L² = [β_sample² − β_instrumental²] (2)

Further, in doped samples of the present kind, the peak broadening owing to the strain is not completely ruled out. The strain and strain-induced peak broadening are related as given in the following eqn (3).

Scherrer equation is reliant on diffraction angle θ as 1/cos θ but not as tan θ as in Williamson–Hall (W–H) method. It implies that Scherrer formula considers only the instrumental broadening but not the strain-induced broadening.

The Williamson–Hall (W–H) equation assumes that the contribution of strain and crystallite size to peak broadening are independent to each other. Therefore the W–H equation is a simple sum of eqn (1) and (3) which is given by

β_hkl = β_s + β_L (4)

The above equation can be rearranged in the form given below

We have employed W–H method (eqn (6)) to perform both crystallite size and strain analysis depending on different θ positions. The W–H plots are depicted in Fig. 3 for all Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) samples that are drawn with 4 sin θ along x-axis and β_hkl cos θ along y-axis. The crystallite size was calculated from the y-intercept and the strain was derived from the slope of the linear fit to the data, respectively. The W–H method assumes the strain to be isotropic and identical along all crystallographic directions. The crystallite size and strain for different compositions of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) are illustrated in Table 3.

Fig. 3 Williamson–Hall plots for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

Table 3 Goldschmidt tolerance factor, crystallite size, lattice strain obtained for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05). Characterization tool: XRD

Name of the compound	Dopant concentration	Crystallite size (nm)	Strain	Tolerance factor
YSFTO	x = 0.05, y = 0.05	112.3	0.273	0.863
	x = 0.10, y = 0.05	119.6	0.284	0.865
	x = 0.15, y = 0.05	124.2	0.317	0.868
YGFTO	x = 0.05, y = 0.05	104.3	0.254	0.859
	x = 0.10, y = 0.05	115.8	0.281	0.861
	x = 0.15, y = 0.05	121.7	0.307	0.859

---

Analysis of crystal structure is considered to be incomplete until it is substantiated by fitting with various crystallographic parameters and multiple pattern variables with a chosen structural model from database. Rietveld proposed a method wherein the individual intensities at each step of powder diffraction data was fit. In this method, a least square technique was incorporated till the best fit is attained between observed powder diffraction data and calculated pattern from the selected model in the database. Rietveld refinement was incorporated to ascertain the validity of orthorhombic structural model and to precisely calculate the lattice parameters. The Rietveld refined plots are depicted in Fig. 4. The Rietveld refinement was done using High score plus software. The experimental data are depicted as transparent rings, while the calculated intensities are depicted as solid lines. The Bragg positions for ‘Pmna’ space group are depicted by the vertical lines at the bottom. The disparity between observed and measured intensities are seen in the form of solid lines in all the figures. The background is corrected using pseudo-voigt function. Parameters like scale factors, shape parameters, isothermal parameters, occupancies and lattice constants were taken to be fixed during the refinement process. We considered oxygen positions to be free parameters whereas other atomic positions were taken to be fixed. The R_wp, weighted profile residual factor which accounts for the background, the R_p, profile residual factor that only takes peaks into account without accounting the background and the goodness of fit (GOF) which describes how well it fits into a set of observations, were found to be low, suggesting the goodness of refinement. The refined lattice parameters and agreement factors are given Table 1.

Fig. 4 Rietveld refined plots for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

Raman studies

Generally, ABO₃ perovskite has a cubic structure with corner-linked BO₆ octahedra with B cations at the centre and A cations in the space between the octahedral, respectively. Raman scattering is forbidden due to symmetry in cubic perovskite structure. From the XRD studies, it is evidenced that samples with different configurations under study are typical example of perovskite structure (ABO₃) and crystallize in orthorhombic structure with Pmna space group. The antiphase tilt of FeO₆ octahedra gives rise to the distorted orthorhombic structure. This tilting of FeO₆ octahedra necessarily introduces a distortion in the RO₁₂ icosahedron (polyhedron).

As per group theory, the orthorhombic Pmna structure has 24 Raman active modes 7A_g + 5B_1g + 7B_2g + 5B_3g. Out of these modes 12 are first order Raman modes.²² The Raman spectra for the system Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) is presented in the Fig. 5.

Fig. 5 Raman spectra for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

The structural analysis could be done by analyzing the Raman bands in the four different spectral regions.

(i) Region 1: 120–200 cm⁻¹, with two peaks 149 cm⁻¹ and 177 cm⁻¹ which are attributed to the vibrations of yttrium ion. According to harmonic oscillator approximation ω ∼ (k/M)^1/2, where ω is the frequency, k is the spring constant and M is atomic mass, the heaviest atom is predicted to vibrate in low wavenumber region and this justifies the shift in the peaks with different dopants and different doping concentrations.

(ii) Region 2: 200–380 cm⁻¹, characterized by 3 different bands. The band at 220 cm⁻¹ is attributed to Fe³⁺ ion vibration and the bands around 270 cm⁻¹ and 340 cm⁻¹ are attributed to Fe³⁺ magnetic ions excitation.

(iii) Region 3: 380–550 cm⁻¹, with two bands of different intensities at 430 cm⁻¹ and 481 cm⁻¹ which are attributed to the excitation of Fe³⁺ magnetic ions.

(iv) Region 4: 550–800 cm⁻¹, characterized by two partially overlapping bands around 600 cm⁻¹, one band is a characteristic of Fe–O bond. The appearance of a second band which is not observed in pristine YFeO₃ might be due to the doping of Ti⁴⁺ at Fe³⁺ sites.

The surface morphology of all the synthesized powdered samples was depicted using backscattered electron micrographs as shown in Fig. 6. All the powder samples were sputtered by platinum to avoid surface charging and were mounted using carbon tape. The samples under study exhibit only one region with uniform colour tonality which is associated with the electron density, proving that all the synthesized samples are monophasic without any impurity phases. The crystallite size distributions are depicted in each micrograph as the insets. The log normal function was used to fit the obtained data.

where 〈D〉 = D₀ exp(σ²/2)

Fig. 6 SEM micrographs of (a) Y_1−xR_xFe_1−(4/3)yTi_yO₃ (x = 0.05, y = 0.05), (b) Y_1−xGd_xFe_1−(4/3)yTi_yO₃ (x = 0.10, y = 0.05), (c) Y_1−xGd_xFe_1−(4/3)yTi_yO₃ (x = 0.15, y = 0.05) (d) Y_1−xSm_xFe_1−(4/3)yTi_yO₃ (x = 0.05, y = 0.05), (e) Y_1−xSm_xFe_1−(4/3)yTi_yO₃ (x = 0.10, y = 0.05), (f) Y_1−xSm_xFe_1−(4/3)yTi_yO₃ (x = 0.15, y = 0.05).

And the mean particle size is given by

σ_D = 〈D〉[exp(σ²) − 1]^1/2

It is observed that the crystallite size increased from 0.28 μm to 0.44 μm for Y_1−xSm_xFe_1−(4/3)yTi_yO₃ (x = 0.05, 0.10, 0.15; y = 0.05) with an error bar of 0.0173 to 0.0241 (Fig. 6a–c) and from 0.49 μm to 0.60 μm for Y_1−xGd_xFe_1−(4/3)yTi_yO₃ (x = 0.05, 0.10, 0.15; y = 0.05) with an error bar of 0.0156 to 0.0210. The reason for such an increase in crystallite size with an increase in dopant concentration could also be due to the reduction of oxygen vacancies. The heterovalent doping of Ti⁴⁺ at Fe³⁺ results in the charge compensation by impregnating such vacancies. This is evidenced by the reduction of pores which can be observed in Fig. 6. Moreover, the partial replacement of Ti⁴⁺ whose ionic radii and atomic weight lower than that of Fe³⁺ enhances the grain growth.²² The crystallite size calculated from XRD analysis varies significantly from that obtained by the SEM analysis. This is ascribed to the agglomeration effects owing to their smaller size and innate magnetism.²³

The energy dispersive X-ray spectra (EDS) of all the synthesized show the peaks associated with Y, Sm, Gd, Fe, Ti and O atoms. The peaks associated with 2.07 keV and 0.27 keV were attributed to Pt and carbon tape (used for sample preparation) respectively. Apart from proving the existence of all the elements associated with the samples under investigation, the EDS spectrum quantifies their relative proportions reasonably well. The lack of elemental impurities in the synthesized samples could be evidenced by the absence of unaccounted peaks (Table 4).

Table 4 Atomic percentage of the different elements present in Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05). Characterization tool: EDS

Name of the Compound	Dopant concentration	Atomic%
		Y L	Sm L	Gd L	Fe K	Ti K	O K
YSFTO	x = 0.05, y = 0.05	17.45	2.21	—	27.04	2.54	50.76
	x = 0.10, y = 0.05	16.81	3.15	—	28.12	2.63	49.29
	x = 0.15, y = 0.05	15.92	4.23	—	28.97	2.74	51.86
YGFTO	x = 0.05, y = 0.05	17.77	—	2.66	26.7	2.69	50.18
	x = 0.10, y = 0.05	16.37	—	3.25	27.4	2.7	50.28
	x = 0.15, y = 0.05	15.99	—	4.1	28.5	2.61	48.8

---

Optical properties

The optical properties unravel the changes that take place in the noticeable Fe d–d transitions which appear in the visible region and uncover the changes that are likely to occur in the bandgap of the material. Diffuse reflectance spectroscopy (DRS) was invoked to study the absorption characteristics and variation in the bandgap with a change in doping concentration. Scattered radiation excluding the specular reflected radiation is collected which matches closely with the Kubelka–Munk function which is given by F(R) = (1 − R)²/2R, where R is the diffuse reflectance.²⁴

The absorbance vs. wavelength plots for all the samples under study are shown in Fig. 7. The Kubelka–Munk function was used to formulate the [F(R)hν]² vs. hν, where h is Planck's constant and ν is the frequency of the source of illumination. The optical band gap energy, E_g, is derived from the intersection of the tangent line at [F(R)hν]² = 0. These plots are shown as the insets in Fig. 7 and the band gap values are given in Table 5.

Fig. 7 Absorption spectra and corresponding Tauc plots for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

Table 5 Energy bandgap values for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05). Characterization tool: UV-Vis

Name of the compound	Dopant concentration	Eg (eV)	Error
YGFTO	x = 0.05, y = 0.05	1.96178	0.03021
	x = 0.10, y = 0.05	1.95268	0.02892
	x = 0.15, y = 0.05	1.913224	0.02756
YSFTO	x = 0.05, y = 0.05	1.930412	0.02791
	x = 0.10, y = 0.05	1.910113	0.03219
	x = 0.15, y = 0.05	1.86448	0.03428

---

In the absorbance plots, there is a strong transition in the range 500–600 nm which corresponds to the electronic transitions implying the charge transfer from valence band of oxygen (O) 2p states to conduction band of iron (Fe) 3d states. These are the allowed p–d transitions.^25,26 The octahedral crystal field splitting of BO₆ octahedra in ABO₃ type perovskite gives rise to splitting of d-orbital into the sub-orbitals of t_2g (d_xy, d_yz, d_xz) and e_g (d_x²−y², d_z²). The intensities of the d–d transitions that occur from t_2g to e_g orbitals were observed to be 10³ times lower than the dipole allowed p–d transitions (O 2p–Fe 3d).²⁷ In the present study, Fe occupies 3+ oxidation state that is to say it is a d⁵ system. The d–d transitions in d⁵ system are forbidden by electric dipole selection rule (Δl = ±1) and spin selection rule (ΔS = 0).²⁷ Even though RFeO₃ samples were dominated by exchange interactions of p–d transitions, we observed the d–d transitions for all the samples under investigation. It is evident from the appearance of intense bands around 1.7 eV in the Tauc plots (Fig. 8).

Fig. 8 d–d transitions from the Tauc plots of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

Tanabe–Sugano (T–S) diagrams were reported to be used to determine the energies of Fe²⁺ (3d⁶), Fe³⁺ (3d⁵), Fe⁴⁺ (3d⁶) systems.²⁸ T–S diagrams envisage the zero spin allowed d–d transitions for Fe³⁺, while they project the spin allowed d–d transitions for Fe²⁺ which is from t⁴_2ge²_g–t³_2ge³_g at 1.277 eV and Fe⁴⁺ which is from t³_2ge¹_g–t²_2ge²_g at 1.72 eV.²⁸ In the present study, the bands around 1.7 eV are attributed to the spin allowed d–d transitions, which comply with the T–S diagrams of spin allowed d–d transitions of Fe⁴⁺. It should be noted from these observations in the absorption spectra of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05), that Fe essentially exists in the mixed states of Fe³⁺ and Fe⁴⁺. This can be further supported by XPS studies. The existence of mixed states of iron can be attributed to the doping effect.

XPS analysis

X-ray photon spectroscopy that provides with a comprehensive estimate of the chemical states of the elements present in the Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) was performed. The deconvolution of Fe 2P_3/2 peak has been done using Shirley background and Gaussian–Lorentzian fitting. The deconvolution gives rise to two different peaks with binding energies around 709.64 eV and 710.72 eV corresponding to Fe³⁺ and Fe⁴⁺ respectively.²⁹ It is observed that Fe⁴⁺ content increased with an increase in dopant concentration at Y³⁺. There is a significant shift in Fe 2P_3/2 peaks towards higher binding energies as depicted in Fig. 9, implying a gradual increase of charge at Fe site.³⁰ This shift is related to the increase of relative amounts of Fe⁴⁺ ions with an increase in dopant concentration. This proves that the Fe exists in mixed oxidation states of Fe³⁺ and Fe⁴⁺. The emergence of Fe⁴⁺ in the synthesized samples is ascribed to the doping of Ti⁴⁺ at Fe³⁺; Sm³⁺ and Gd³⁺ at Y³⁺ sites and presence of oxygen vacancies.

Fig. 9 XPS spectra obtained for Fe 2p_3/2 for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

Two peaks are shown for O 1s XPS pattern. The peak around 528 eV is attributed to lattice oxygen and peak at 532 eV is attributed to the surface adsorbed oxygen.³¹ In case of Sm³⁺ and Gd³⁺ doping with x = 0.05, the peak at 528 eV is not so visible. The reason might be the loss of electron making it O⁻ from O²⁻ ion and giving rise to prominent 532 eV peak, suggesting the higher concentrations of oxygen vacancies. With the increase in dopant concentration at Y³⁺ the height and area of the peak at 528 eV (lattice oxygen peak) gradually increases and becomes more evident. A detailed quantitative analysis was used to evaluate the content of oxygen deficiencies. The oxygen content was separated into lattice oxygen (O²⁻) and surface adsorbed oxygen (O⁻) using deconvolution. We have calculated the atomic ratio of O²⁻/O⁻ which gives an estimate of the concentration of oxygen deficiencies. For Y_1−xSm_xFe_1−(4/3)yTi_yO₃ (x = 0.05, 0.10, 0.15; y = 0.05) the atomic ratio of O²⁻/O⁻ increased from 2.19 to 10.134 whereas for Y_1−xGd_xFe_1−(4/3)yTi_yO₃ (x = 0.05, 0.10, 0.15; y = 0.05) O²⁻/O⁻ ratio increased from 2.52 to 10.70. This increase in atomic ratio of O²⁻/O⁻ with an increase in dopant concentration implies the reduction of concentration of oxygen deficiencies (Fig. 10).

Fig. 10 XPS spectra obtained for O 1s for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

Magnetic properties

The magnetic properties in rare-earth orthoferrites originate from the super-exchange interaction of Fe³⁺–Fe³⁺, R³⁺–R³⁺, R³⁺–O²⁻–Fe³⁺.³² The room temperature magnetization versus applied magnetic field (M–H) loop of YFeO₃ suggests that it is an antiferromagnetic material.¹² The GKA (Goodenough, Kanamori and Anderson) rules predict the antiferromagnetic behaviour of YFeO₃ through oxygen mediated super-exchange interaction of pristine YFeO₃.³³ As mentioned earlier, the d⁵ orbitals of Fe³⁺ split into triply degenerate t_2g (d_xy, d_yz, d_xz) and doubly degenerate e_g (d_x²−y², d_z²) orbitals. The e_g orbitals of Fe³⁺ which are along the crystal axes overlap with 2p orbitals of O²⁻ which give rise to the super-exchange of Fe³⁺–O²⁻–Fe³⁺ at 180°. However, the magnetic moment alignment of Fe³⁺ ions is not precisely antiparallel giving rise to the canting of the spins. This results in a small magnetization effect and weak ferromagnetism in YFeO₃. Literature reports show that there is an obvious effect on magnetic properties with magnetic and diamagnetic partial substitution/doping.^11,12,34 Fig. 11 gives the M–H hysteresis loops for different compositions of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) at room temperature. There is an increase in magnetization with an increase in dopant concentration at Y³⁺. However, even at the highest applied magnetic field of 2 T, the magnetization doesn't seem to saturate. The pinching or linear behaviour of pristine YFeO₃ is a typical anti-ferromagnetic behaviour. The hysteresis behaviour is attributed to the significantly improved canting of spins. The enhancement in magnetization is due to larger spin canting associated with the presence of Ti⁴⁺ dopant ions and Fe³⁺ vacancies that are required for restoration of charge balance in the lattice. Also, the uncompensated surface spins of Fe³⁺ ions can enhance the magnetization. Yet another reason might be the formation of Y³⁺–O²⁻–R³⁺ instead of R³⁺–O²⁻–R³⁺ chain due to large magnetic moment of the dopant ions Gd³⁺, Sm³⁺. It is also observed that there a significant enhancement in coercivity. The possible reasons for this might be the formation of frustrated spin systems arising as a result of non-collinear magnetic ordering originating from the presence of uncompensated spins due to partial substitution of Ti⁴⁺ dopant ions at Fe³⁺ sites and balancing of Fe³⁺ vacancies.

Fig. 11 M-H loops for Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

Lattice defects could be introduced by the doping of different radii ions at Y³⁺ site. Host lattice defects that are intrinsic can also be activated when different ions partially occupy Y³⁺ or Fe³⁺ sites which is evident from the increase in lattice constants. The increase in lattice constants with the increase in dopant concentration results in the increase in exchange interaction leading to the enhancement in magnetization. In the scenario of Ti⁴⁺ doping at Fe³⁺, the ionic radius of Ti⁴⁺ is almost similar to that of Fe³⁺, a strong surface anisotropy plays the major role in the significant increase of coercive field. The enhancement in H_c (coercivity) also suggests the possible existence of ferromagnetic (FM) exchange interaction.

It should be noted here that Fe–O–Fe super-exchange coupling is a predominant factor in the determination of magnetic ordering. This super-exchange interaction could be influenced by valence states of magnetic ions. Hence, it becomes necessary to have information about valence states of Fe to understand the origin of magnetic behaviour comprehensively. XPS analysis revealed that Fe exists in Fe³⁺ and Fe⁴⁺ mixed states. The XPS studies revealed that the relative amount Fe⁴⁺ increases with an increase in dopant concentration. It can also be speculated that the amount of Fe⁴⁺ increases with a decrease in oxygen deficiencies. This view is supported by our analysis of O 1s XPS pattern in the earlier section. Previous studies have shown that the 180° super-exchange coupling of Fe⁴⁺–O²⁻–Fe⁴⁺ is ferromagnetic.³⁵ Therefore, the enhancement in ferromagnetic ordering can be due to Fe⁴⁺–O²⁻–Fe⁴⁺ super-exchange coupling. However, the presence of Fe³⁺ ions and oxygen deficiencies can induce a different kind of magnetic order which contributes to a small magnetization.³⁶ It is well known that even a minute structural change/distortion such as tilting of FeO₆ octahedra impact magnetic properties largely. In particular, parameters such as bond angle and bond length which determine orbital overlap, charge transfer and exchange interaction will have a potential effect on magnetic properties. The XPS and VSM analyses carried out on the samples under investigation prompted us to propose two types of exchange coupling interactions i.e., antiferromagnetic super-exchange coupling from Fe³⁺–O²⁻–Fe³⁺ and Fe⁴⁺–O²⁻–Fe⁴⁺ and double exchange coupling from Fe³⁺–O²⁻–Fe⁴⁺. Also the magnetic properties are affected by oxygen vacancies which interrupt the bridges between O and Fe/Ti and are undesirable from the point of hopping. The detailed quantitative information on the crystallographic changes obtained by XRD and oxygen vacancies allows us to comprehend the magnetic properties correlated by exchange coupling mechanism. We have observed an increase in Fe/Ti–O distance with an increase in dopant (Sm/Gd) concentration, resulting in a decrease in O 2p bandwidth. This decreases the orbital overlap between Fe (3d) and O (2p).

In Fe³⁺–O²⁻–Fe³⁺ super-exchange coupling interaction, the charge transfer mainly depends on overlap between Fe (3d) and O (2p) and is maximum when the distortion angle is zero. Such kind of super-exchange leads to anti-ferromagnetism. Since all the prepared samples are in orthorhombic symmetry which is evidenced from Goldschmidt tolerance factor, have the largest deviation from the ideal cubic structure. As the distortion increases with an increase in dopant (Sm³⁺/Gd³⁺) concentration listed in the Table 2, the behavior of the samples deviates from antiferromagnetic to ferromagnetic.

In Fe³⁺–O²⁻–Fe⁴⁺, the conducting electrons hop from Fe³⁺ towards Fe⁴⁺ via O²⁻ bridge is a double exchange coupling interaction. As mentioned earlier, the increase in dopant (Sm³⁺/Gd³⁺) concentration at Y³⁺ sites resulted in the increase in Fe/Ti–O distance which consequently reduces the mobility of double exchange electrons which decrease the ferromagnetic property of the prepared samples. However, the reduction in oxygen vacancies evidenced from XPS studies paves the way for the exchange coupling mechanism to take place effectively thus resulting in the improved ferromagnetic behavior of the prepared samples. At this juncture, it is worth mentioning that we did not consider Ti⁴⁺–O²⁻–Ti⁴⁺ exchange interaction since Ti⁴⁺ as such is non-magnetic, the doping is comparatively minute and the probability of Ti atoms next to each other is negligibly small considering the random distribution of ions in Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05). Table 6 elucidates the maximum magnetization (M_m), remanent magnetization (M_r) and coercivity (H_c) of Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05).

Table 6 The maximum magnetization (M_m), remanent magnetization (M_r) and coercivity (H_c). Characterization tool: VSM

Name of the compound	Dopant concentration	Mm (10^−3) emu g^−1	Mr (10^−3) emu g^−1	Hc (coercive field) kOe
YSFTO	x = 0.05, y = 0.05	14.02	4.62	10.973
	x = 0.10, y = 0.05	18.06	7.15	11.162
	x = 0.15, y = 0.05	24.27	11.14	12.784
YGFTO	x = 0.05, y = 0.05	12.39	3.76	6.551
	x = 0.10, y = 0.05	20.47	4.28	9.24
	x = 0.15, y = 0.05	22.84	6.26	11.168

---

Conclusions

Phase-pure Y_1−xR_xFe_1−(4/3)yTi_yO₃ (R = Sm, Gd; x = 0.05, 0.10, 0.15; y = 0.05) were synthesized via low-temperature solid-state synthesis route. Thus prepared powders were found to be orthorhombic matching to that of parent compound YFeO₃. The optical bandgap for all the samples under investigation was observed to decrease with the increase in dopant concentration. The enhancement in magnetization and coercivity suggests a transition from antiferromagnetic to ferromagnetic behaviour. The improved magnetic properties are attributed to the existence of Fe⁴⁺ along with Fe³⁺ and the reduction of oxygen deficiencies.

Funding

This research work did not obtain any particular grants from state, private or non-profit funding agencies.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors devote the research to the founder chancellor Bhagawan Sri Sathya Sai Baba and thank the Sri Sathya Sai Central Trust for providing synthesis and characterization facilities at Central Research Instruments Facility (SSSIHL-CRIF). The authors also thank Japan Advanced Institute of Science and Technology, Japan for providing with XPS characterization facility.

References

1. N. A. Spaldin, MRS Bull., 2017, 42, 385–389.
2. S. Cao, H. Zhao, B. Kang, J. Zhang and W. Ren, Sci. Rep., 2015, 4, 5960.
3. S. M. Yusuf, A. Kumar and J. V. Yakhmi, Appl. Phys. Lett., 2009, 95, 11–14.
4. J. Mao, Y. Sui, X. Zhang, Y. Su, X. Wang, Z. Liu, Y. Wang, R. Zhu, Y. Wang, W. Liu and J. Tang, Appl. Phys. Lett., 2011, 98, 2009–2012.
5. C. W. Nan, M. I. Bichurin, S. Dong, D. Viehland and G. Srinivasan, J. Appl. Phys., 2008, 103, 031101.
6. L. Mitoseriu, Bol. Soc. Esp. Ceram. Vidrio, 2005, 44, 177–184.
7. H. Schmid, J. Phys.: Condens. Matter, 2008, 20, 434201.
8. G. King and P. M. Woodward, J. Mater. Chem., 2010, 20, 5785–5796.
9. N. Singh, J. Y. Rhee and S. Auluck, J. Korean Phys. Soc., 2008, 53, 806–811.
10. D. Van Tac, V. O. Mittova and I. Y. Mittova, Inorg. Mater., 2011, 47, 521–526.
11. X. Yuan, Y. Sun and M. Xu, J. Solid State Chem., 2012, 196, 362–366.
12. P. S. J. Bharadwaj, S. Kundu, K. Vijay Sai and K. B. R. Varma, J. Phys.: Condens. Matter, 2020, 32, 035810.
13. M. Wang and T. Wang, Materials, 2019, 12, 1–11.
14. N. O. Khalifa, H. M. Widatallah, A. M. Gismelseed, F. N. Al-Mabsali, R. G. S. Sofin and M. Pekala, Hyperfine Interact., 2016, 237, 46.
15. M. C. Carotta, G. Martinelli, Y. Sadaoka, P. Nunziante and E. Traversa, Sens. Actuators, B, 2002, 48, 270–276.
16. J. Ding, X. Lü, H. Shu, J. Xie and H. Zhang, Mater. Sci. Eng. B Solid State Mater. Adv. Technol., 2010, 171, 31–34.
17. R. L. Zhang, C. Le Chen, K. X. Jin, L. W. Niu, H. Xing and B. C. Luo, J. Electroceram., 2014, 32, 187–191.
18. M. Siemons, T. Weirich, J. Mayer and U. Simon, Z. Anorg. Allg. Chem., 2004, 630, 2083–2089.
19. Z. Zhou, L. Guo, H. Yang, Q. Liu and F. Ye, J. Alloys Compd., 2014, 583, 21–31.
20. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
21. M. O'Keeffe and B. G. Hyde, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1977, 33, 3802–3813.
22. E. VenkataRamana, F. Figueiras, A. Mahajan, D. M. Tobaldi, B. F. O. Costa, M. P. F. Graca and M. A. Valente, J. Mater. Chem. C, 2016, 4, 1066–1079.
23. A. Shokri, S. Farjami and K. Boustani, Ceram. Int., 2018, 44, 22092–22101.
24. P. Kubelka, J. Opt. Soc. Am., 1948, 38, 448–457.
25. V. Gopalan, S. W. Cheong, J. L. Musfeldt, N. High and M. Field, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 4–7.
26. J. H. Jung, M. Matsubara, T. Arima, J. P. He, Y. Kaneko and Y. Tokura, Phys. Rev. Lett., 2004, 93, 1–4.
27. A. S. Moskvin, A. A. Makhnev, L. V. Nomerovannaya, N. N. Loshkareva and A. M. Balbashov, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 82, 035106.
28. Y. Tanabe, J. Phys. Soc. Jpn., 1956, 11, 864–877.
29. T. Matsui, E. Taketani, R. Sato and K. Morii, J. Phys. D: Appl. Phys., 2007, 40, 6066–6070.
30. M. Asim Farid, H. Zhang, X. Lin, A. Yang, S. Yand, G. Li, F. Liao and J. Lin, RSC Adv., 2015, 5, 12866–12871.
31. H. Aono, M. Sato, E. Traversa and M. Sakamoto, J. Am. Ceram. Soc., 2001, 47, 341–347.
32. A. Jaiswal, R. Das, S. Adyanthaya and P. Poddar, J. Phys. Chem. C, 2011, 115, 2954–2960.
33. P. W. Anderson, Phys. Rev., 1959, 115, 2–13.
34. K. T. Jacob, G. Rajitha and N. Dasgupta, Indian J. Eng. Mater. Sci., 2012, 19, 47–53.
35. S. Mori, J. Phys. Soc. Jpn., 1970, 28, 44–50.
36. T. Matsui, H. Tanaka, N. Fujimura, T. Ito, H. Mabuchi and K. Morii, Appl. Phys. Lett., 2002, 81, 2764–2766.

---